0712.0026/ms.tex
1: \documentclass{emulateapj}
2: 
3: \newcommand{\ntwo}{N$_{2}$H$^{+}$}
4: \newcommand{\ntwoi}{N$_{2}$H$^{+}$ 1$_{01}$--0$_{12}$}
5: \newcommand{\hcop}{HCO$^{+}$}
6: \newcommand{\meth}{CH$_{3}$OH}
7: \newcommand{\cs}{C$^{34}$S}
8: \newcommand{\htcop}{H$^{13}$CO$^{+}$}
9: \newcommand{\hcn}{H$^{13}$CN}
10: \newcommand{\iras}{IRAS~23033}
11: \newcommand{\msun}{M$_{\odot}$}
12: \newcommand{\um}{$\mu$m}
13: \newcommand{\kms}{km~s$^{-1}$}
14: \newcommand{\sig}{$\sigma$}
15: \newcommand{\cmt}{cm$^{-3}$}
16: \newcommand{\cmtwo}{cm$^{-2}$}
17: \newcommand{\gcmt}{g~cm$^{-3}$}
18: \newcommand{\gcmtwo}{g~cm$^{-2}$}
19: \newcommand{\mjb}{mJy~beam$^{-1}$}
20: \newcommand{\mjbkms}{mJy~beam$^{-1}$~km~s$^{-1}$}
21: 
22: 
23: \begin{document}
24: 
25: \title{Deconstructing the High-Mass Star-Forming Region IRAS 23033+5951}
26: 
27: \author{Michael A. Reid}
28: \affil{Harvard-Smithsonian Center for Astrophysics, Submillimeter Array 
29: Project, 645 North A'ohoku Pl., Hilo, HI, 96720, USA}
30: 
31: \author{Brenda C. Matthews} 
32: \affil{Herzberg Institute of Astrophysics, National Research Council of 
33: Canada, Victoria, BC V9E 2E7, Canada}
34: 
35: \begin{abstract}
36: 
37:         We report interferometric observations of the high-mass 
38: star-forming object IRAS 23033+5951.  Our observations reveal two massive 
39: molecular cloud cores, designated IRAS 23033+5951-MMS1 and IRAS 
40: 23033+5951-MMS2.  MMS1 has already formed a massive protostar 
41: and MMS2 appears to be on the verge of doing so.  The latter 
42: core may be an example of a massive analogue to a ``Class 0'' star-forming 
43: object.  The more evolved core shows some evidence of \ntwo\ destruction
44: near the protostar, consistent with similar findings in low-mass 
45: star-forming objects.  In addition to the already-known prominent \hcop\ 
46: outflow, our SiO 2--1, and \meth\ 2--1 maps show evidence for two more 
47: candidate outflows, both presumably less powerful than the main one.  
48: Both cores are embedded in an elongated feature whose major axis is 
49: oriented almost exactly perpendicular to the axis of the most prominent 
50: outflow in the region.  Although it has many of the characteristics of a 
51: disk, the 87,000~AU (0.42~pc) diameter of this structure suggests that 
52: it is more likely to be the flattened, rotating remnant of the natal 
53: molecular cloud fragment from which the star-forming cores condensed.  We 
54: conclude that IRAS 
55: 23033+5951 is an excellent example of massive star formation proceeding in 
56: relative isolation, perhaps by the method of monolithic collapse and disk 
57: accretion.
58: 
59: \end{abstract}
60: 
61: \keywords{stars: formation --- ISM: individual (IRAS 23033+5951) --- ISM: 
62: molecules --- ISM: kinematics and dynamics}
63: 
64: \section{Introduction}
65: 
66: 	Two theories of massive star formation dominate the literature 
67: (see \citealt{mo07} for a recent comprehensive review).  
68: According to the first, massive stars form by the gravitational collapse 
69: of molecular cloud cores, which have masses of tens to hundreds of solar 
70: masses.  The molecular cloud cores are themselves produced by the 
71: fragmentation of molecular cloud clumps, which have masses of hundreds to 
72: thousands of solar masses.  In this ``monolithic collapse'' model, massive 
73: cores are envisioned to collapse to form single massive stars or small 
74: multiple systems via a process schematically similar to that which forms 
75: low-mass stars \citep{mt03,kmk05,k06}.  The other prominent theory of 
76: massive star formation, known as competitive accretion, holds that all 
77: stars begin as low-mass protostars which then compete to accrete some 
78: fraction of the mass of their parent molecular cloud clump, sometimes 
79: coalescing in the process \citep{bbz98,b01a,b01b,bvb04}.  In the 
80: competitive accretion scenario, massive stars are simply those which 
81: accrete most successfully.  It is hard to imagine that competitive 
82: accretion does not occur at some level in all clustered star-forming 
83: regions.  However, recent theoretical work suggests it is not the dominant 
84: mode by which massive stars form \citep{kmk05}.  Direct evidence capable 
85: of distinguishing these modes of star formation remains elusive.  To prove 
86: that massive stars can form by the collapse of individual molecular cloud 
87: cores would require several conditions to be verified by direct 
88: observation: (a) a massive protostar forming inside a well-defined 
89: molecular cloud core, (b) an infall signature in that molecular cloud core 
90: which is consistent with a high rate of protostellar accretion, (c) a 
91: massive circumstellar disk (whose properties are not yet well 
92: constrained), and (d) a collimated outflow with a high entrained mass.
93: 
94: 	Relatively few massive star-forming objects with most or all of 
95: these characteristics have been identified.  Perhaps the best example is 
96: IRAS~20216+4104 \citep{ces97,ces99}, for which there is strong evidence of 
97: a collapsing Keplerian disk around a massive protostar.  Other examples, 
98: such as Cepheus~A~HW2 \citep{patel05}, are hotly disputed.  The 
99: principal obstacle in the study of massive protostellar systems is lack of 
100: spatial resolution.  Because massive protostars typically lie several 
101: kiloparsecs or more from the Sun, it is extremely challenging to spatially 
102: resolve their accretion disks, even with modern millimeter and 
103: submillimeter interferometers.  Our inadequate understanding of the 
104: chemistry of massive protostellar systems poses another significant 
105: challenge, making it difficult to select appropriate molecular transitions 
106: with which to study the kinematics of such systems.
107: 
108: 
109: 	IRAS 23033+5951 (\iras\ hereafter) is an active star-forming 
110: region lying at a distance of 3.5~kpc in the Cepheus star-forming complex.  
111: Evidence of ongoing massive star formation has been detected in the form 
112: of compact radio continuum emission \citep{srid02,beu02b}, maser emission 
113: \citep{beu02b}, and two discrete dust continuum sources at 1.2~mm 
114: \citep{beu02}.  However, in contrast to many high-mass star-forming 
115: regions, relatively few discrete sources have been detected and there is 
116: evidence for only one substantial outflow \citep{beu02c,mz03}.  Thus, 
117: \iras\ provides an attractive opportunity to study massive star formation 
118: in a relatively quiescent, morphologically simple environment in which the 
119: usual complicating factors---crowding, abundant contamination from radio 
120: continuum emission, lack of spatial resolution---are mitigated.
121: 
122: 	Our goal in studying \iras\ was to characterize its kinematics and 
123: determine whether it is a candidate example of a massive star-forming by 
124: monolithic collapse, as suggested by its simple morphology and known 
125: kinematics.  We observed \iras\ in several molecular transitions using the 
126: Berkeley-Illinois-Maryland Association (BIMA) array.  Each spectral line 
127: was selected to probe a particular kinematic aspect of \iras\, including 
128: outflows, rotation, infall, and turbulence.  In \S\ref{sec:obs}, we 
129: discuss the observations and the data reduction procedure.  In 
130: \S\ref{sec:morph}, we discuss the general morphology and evolutionary 
131: state of \iras, as determined from both the continuum and molecular line 
132: observations.  In \S\ref{sec:spec}, we concentrate on the kinematic 
133: results derived from the spectral line observations.  Our results are 
134: summarized in \S\ref{sec:summary}.
135: 
136: 
137: \section{Observations and Data Reduction}
138: \label{sec:obs}
139: 
140: \begin{deluxetable*}{llcccc}
141: \tablecaption{Molecular Lines Observed\label{tab:tunings}}
142: \tablewidth{0pt}
143: \tablehead{\colhead{Frequency} & \colhead{Molecular} & \colhead{Array} & \colhead{k$\lambda$} & \colhead{Velocity Res.} & \colhead{rms/channel}\\ \colhead{(GHz)} & \colhead{Line} & \colhead{Configurations} & \colhead{Range} & \colhead{(\kms)} & \colhead{(Jy beam$^{-1}$)} }
144: \startdata
145: 86.243440 & SiO 2--1 & B, C & 2--69 & 0.337 & 0.13 \\
146: 86.340167 & \hcn\ 1--0 & B, C, D & 2--68 & 0.085 & 0.21\\
147: 86.754330 & \htcop\ 1--0 & B, C & 2--69 & 0.337 & 0.13 \\
148: 89.188526 & \hcop\ 1--0 & C, D & 2--25 & 0.041 & 0.75 \\
149: 93.173777 & \ntwo\ 1--0 & B, C & 2--74 & 0.079 & 0.079\\
150: 96.412961 & \cs\ 2--1 & B, C & 2--74 & 1.21 & 0.058 \\
151: 96.741377 & \meth\ 2--1 & B, C & 2--74 & 0.302 & 0.12\\
152: \enddata
153: \end{deluxetable*}
154: 
155: 	Observations were made over 8 nights between 2003 October and 2004 
156: May using the BIMA interferometer in Hat Creek, CA.  All of the 
157: observations were made using single pointings toward the position $\alpha 
158: = 23^{\rm h}05^{\rm m}25\fs7$, $\delta = +60\degr08\arcmin08\arcsec$ 
159: (J2000).  Tracks were obtained in the B, C, and D configurations of the 
160: array and included tunings at 86, 89, and 93~GHz.  Not all of the spectral 
161: lines were observed with the same velocity resolution during every track.  
162: Table~\ref{tab:tunings} provides a complete summary of the lines observed, 
163: array configurations used, and the spectral line velocity resolutions used 
164: in most of our analysis.
165: 
166: 	The data were flagged, calibrated, CLEANed and analyzed using the 
167: MIRIAD software package \citep{miriad}.  Observations of the nearby 
168: quasars 0102+584 and 2203+317 were made every 30 minutes during the 
169: observations and were used to correct instrumental and atmospheric phase 
170: and amplitude variations in the data.  Absolute flux calibration was 
171: performed using observations of Uranus taken during each track.  Based on 
172: the variations in the flux of the quasar from one epoch to the next, we 
173: estimate the accuracy of our flux calibration to be approximately 30\%.  
174: In order to maximize the signal-to-noise ratio of the data, we typically 
175: used natural weighting of the visibilities in generating our images. Where 
176: the analysis required higher spatial resolution than that provided by 
177: natural weighting, some of the data were processed using a MIRIAD 
178: ``robust'' parameter of 0.5, at the cost of a lower signal-to-noise ratio.
179: 
180: 	Our 3~mm continuum map was produced by combining continuum data 
181: from all of the different tunings around 3~mm.  The resulting CLEANed, 
182: naturally weighted image has a beam size of 5\farcs5$\times$4\farcs8 and 
183: an rms sensitivity of 2.9~\mjb.
184: 
185: \section{Morphology: Multiple Massive Cores}
186: \label{sec:morph}
187: 
188: 	The initial millimeter continuum observations of \iras\ with the 
189: IRAM 30~m telescope showed a singly peaked structure \citep{beu02}.  
190: Subsequent BIMA array observations of the source at 2.6~mm showed two 
191: continuum peaks \citep{beu02b}.  Our BIMA array 3~mm continuum image, 
192: shown in Figure~\ref{fig:contmodel}, reveals \iras\ to be composed of at 
193: least three continuum peaks joined by contiguous emission, plus a weak 
194: fourth peak visible to the east at the 3~\sig\ level.  We designate these 
195: three millimeter continuum peaks as IRAS~23033+5951-MMS1, 
196: IRAS~23033+5951-MMS2, and IRAS~23033+5951-MMS3 (hereafter, MMS1, MMS2, and 
197: MMS3) as shown in Figure~\ref{fig:contmodel}.  We used a simple model to 
198: determine the fraction of the continuum emission which is associated with 
199: each of the continuum peaks.  The model, which is shown in the middle 
200: panel of Figure~\ref{fig:contmodel}, fits the two bright peaks (MMS1 and 
201: MMS2) with an elliptical Gaussian and the third contiguous peak (MMS3) 
202: with a point source.  After model subtraction, the residuals in the region 
203: of the three main peaks are below 2~\sig, as shown in the right panel of 
204: Figure~\ref{fig:contmodel}.
205: 
206: 	To compute the total masses of each of these continuum sources, 
207: we follow the prescription of \citet{h83}:
208: \begin{equation}
209: \label{eq:fluxtomass}
210: {\rm M} =
211: \frac{S_{\rm int}d^{2}}{\kappa B_{\lambda}({\rm T}_{\rm dust})}~~,
212: \end{equation}
213: \noindent where M, d, $S_{\rm int}$, $\kappa$, ${\rm T}_{\rm dust}$ are 
214: the 
215: mass, distance, 
216: integrated flux, dust opacity per unit mass column density, and dust 
217: temperature of the source, respectively.  $B_{\lambda}({\rm T}_{\rm 
218: dust})$ is the Planck function at wavelength $\lambda$ and temperature ${\rm T}_{\rm dust}$.  We assume ${\rm T}_{\rm dust}=30$~K, $d=3.5$~kpc 
219: \citep{harju}, a dust emissivity index of
220:  $\beta =1.5$, 
221: $\kappa = 0.1(250.0~\mu{\rm m}/\lambda)^{\beta}$~g$^{-1}$~cm$^{2}$ 
222: \citep{h83}, and a gas-to-dust mass ratio of 100.  Under these 
223: assumptions, we derive total masses of 225~\msun\ for MMS1, 205~\msun\ for 
224: MMS2, and 51~\msun\ for MMS3.  
225: 
226: 	\citet{wil04} observed \iras\ using the JCMT and measured 850~\um\ 
227: and 450~\um\ integrated fluxes of 10.4$\pm$0.2~Jy and 48.6$\pm$2.6~Jy, 
228: respectively, for a region encompassing MMS1, MMS2, and MMS3.  Using these 
229: fluxes and assuming $\beta = 0.9$ and T$_{\rm dust}$ = 52~K, \citet{wil04} 
230: compute masses of 237~\msun\ at 850~\um\ and 106~\msun\ at 450~\um\ for 
231: the \iras\ complex.  Similarly, using single-dish measurements at 1.2~mm, 
232: \citet{beu02} derived a mass of 2327~\msun\ for \iras, assuming $\beta =2$ 
233: and T$_{\rm dust}$ = 52~K.  The differences between the calculated masses 
234: are caused by three factors: the differences in the measured fluxes, the 
235: differences in the assumed dust opacities, and the differences in the 
236: assumed dust temperatures.  Our interferometric maps filter out much of 
237: the flux visible in the single-dish maps.  Because our measurements were 
238: taken at a different wavelength than any of the single-dish measurements, 
239: the computation of the exact amount of flux filtered out by the 
240: interferometer is dependent on the assumed dust opacity and temperature.
241: 
242: \begin{figure*}
243: \begin{center}
244: \includegraphics[width=0.8\columnwidth,angle=270]{f1.eps}
245: \caption{Continuum emission at 3~mm (\emph{left}), a model of the 
246: continuum emission consisting of 2 elliptical gaussians and a point source 
247: (\emph{middle}), and the residuals obtained after subtracting the model 
248: from the data (\emph{right}).  The contours start at 3$\sigma$ in the two 
249: left panels and 2$\sigma$ in the right panel; the interval between 
250: contours is 1$\sigma$ in all three panels, where $\sigma = 2.9$~\mjb.
251: The 
252: ellipse in the lower right of the left panel shows 
253: the size and orientation of the synthesized beam.  
254: The star, triangle, and cross with error ellipse 
255: indicate the positions of the MSX source, H$_{2}$O 
256: maser \citep{beu02b}, and IRAS point source, 
257: respectively. \label{fig:contmodel}} 
258: \end{center} \end{figure*}
259: 
260: 
261: 	If we adopt the same flux-to-mass conversion prescription as 
262: \citet{wil04} (namely, assuming T$_{\rm dust}$ = 52~K and using the 
263: \citet{oss94} dust opacities extrapolated to 3.3~mm) we calculate masses 
264: of 131~\msun, 120~\msun, and 29.5~\msun for MMS1, MMS2, and MMS3, 
265: respectively.  We note, however, that the \citet{oss94} opacities were not 
266: originally tabulated for wavelengths as long as 3.3~mm, nor were they 
267: specifically meant to apply to massive star-forming regions.  Similarly, 
268: if we use the prescription of \citet{beu02} for converting integrated flux 
269: to mass, we find masses of 459~\msun, 419~\msun, and 103~\msun\ for MMS1, 
270: MMS2, and MMS3, respectively.  Clearly the mass uncertainties are quite 
271: large.  Nevertheless, even allowing for generous uncertainties in the 
272: flux-to-mass conversion and in the star formation efficiency, at least 
273: MMS1 and MMS2 appear to have enough mass to each form one or more massive 
274: stars.
275: 
276: 	Gravitationally bound cores capable of forming massive stars 
277: typically have very high densities, $n \geq 10^5$~\cmt, and tend to have 
278: virial ratios, $\alpha_{\rm vir} = M_{\rm vir}/M$, close to unity 
279: \citep{p97,mt03}.  Do either of the spatially resolved cores in \iras\ 
280: meet these criteria?  We can compute the mean densities of the two 
281: brightest cores using their total masses (assumed to be 
282: 225~\msun\ for MMS1 and 205~\msun\ for MMS2, as computed above) and their 
283: mean radii, defined as half of the mean of their major and minor axes, as 
284: determined from the fits shown in Fig~\ref{fig:contmodel}.  These mean 
285: radii are $0.14\pm0.02$~pc for MMS1 and $0.15\pm0.03$~pc for MMS2.  In 
286: this way, we compute mean gas densities of ($3\pm1$)~$\times~10^6$~\cmt\ 
287: and ($2\pm1$)~$\times~10^6$~\cmt, for MMS1 and MMS2, respectively.  Such 
288: high gas densities again qualify both cores as potential sites of massive 
289: star formation.
290: 	
291: 	Figure~\ref{fig:disklines} shows integrated intensity maps of 
292: \iras\ in four molecular lines which principally trace dense gas: \ntwo\ 
293: 1--0, \htcop\ 1--0, \hcn\ 1--0, and \cs\ 2--1.  These four dense gas 
294: tracers peak at different locations.  The best column density tracers 
295: should be \ntwo\ 1--0 and dust, because both of these should be optically 
296: thin and relatively (although perhaps not entirely) unaffected by the 
297: complex carbon chemistry which can change the abundances of \htcop\ and 
298: \hcn\.  There is no indication in the \ntwoi\ line shape that this 
299: transition is optically thick.  Indeed, comparison of 
300: Figure~\ref{fig:contmodel} and Figure~\ref{fig:disklines} shows that the 
301: spatial distribution of \ntwo\ is very similar to the distribution of 
302: dust.  The peak positions of \ntwo\ and dust are separated by less than a 
303: beam width in both MMS1 and MMS2 (see also Fig.~\ref{fig:peaks}).  Hence, 
304: we conclude that, as expected, \ntwo\ 1--0 and dust both trace the densest 
305: gas in the region.  We can therefore use the \ntwoi\ line widths and the 
306: masses determined from the continuum emission to compute the virial ratios 
307: for MMS1 and MMS2.  We derive the \ntwo\ line widths from the isolated 
308: (i.e. non-blended) hyperfine component of the \ntwo\ line.  The \ntwoi\ 
309: line widths (full-width, half-power) in MMS1 and MMS2 are $2.5\pm0.3$ and 
310: $2.4\pm0.2$ \kms, respectively.  If we approximate the cores as spheres 
311: with uniform densities equal to the mean densities stated above and assume 
312: that the shape of the \ntwoi\ line is approximately Gaussian, then the 
313: virial masses of the cores are given by
314: 
315: \begin{equation}
316: M_{\rm vir} = \frac{5R(\Delta v)^{2}}{8{\rm G}\ln{2}}~~,
317: \label{eq:mvir}
318: \end{equation}
319: 
320: \noindent where $R$ is the radius of the core and $\Delta v$ is the 
321: observed \ntwo\ line width.  By this prescription, the virial masses of 
322: MMS1 and MMS2 are 178 and 184 \msun\ respectively.  Hence, their virial 
323: ratios are 0.79 and 0.90.  The virial parameter can be interpreted as a 
324: measure of a cloud's degree of gravitational binding, with values less 
325: than 1 indicating a gravitationally bound cloud.  Therefore, subject to 
326: the uncertainties in determining both the virial and dust continuum masses 
327: of the \iras\ cores, we conclude that they are both gravitationally bound.  
328: This interpretation is consistent with the fact that MMS1 is clearly 
329: forming at least one star.  It also suggests that MMS2 might already be 
330: forming a star, or may eventually do so.
331: 
332: \begin{figure*} 
333: \begin{center} 
334: \includegraphics[angle=270]{f2.eps} 
335: \caption{Integrated intensity of four lines in \iras.  Counter-clockwise 
336: from the top left panel, the lines are: \ntwo, \htcop, \cs, and \hcn.  The 
337: contours start at 3\sig\ in all panels 
338: and increase by steps of 4, 1, 1, and 1\sig, respectively.  The \ntwo\ 
339: intensity was integrated over all seven hyperfine components of the 
340: line.  The sizes and 
341: orientations of the four beams are indicated by the ellipses in the lower 
342: right corners of each panel.  White crosses indicate the positions of the 
343: 3~mm continuum peaks from Fig.~\ref{fig:contmodel}.  The 1~\sig\ rms in 
344: the panels are, from left to right, 290~\mjbkms, 210~\mjbkms, 150~\mjbkms, 
345: and 320~\mjbkms.	
346: \label{fig:disklines}}
347: \end{center}
348: \end{figure*}
349: 
350: \begin{figure}
351: \includegraphics[width=0.8\columnwidth,angle=270]{f3_bw.eps}
352: \caption{Molecular line emission peaks (\emph{symbols}) overlaid on the 
353: 3~mm continuum map (\emph{greyscale and contours}).  The meaning of each 
354: symbol is indicated in the legend.  For simplicity, the 
355: positional error bars are not shown; they are typically 
356: $\pm$1--2\arcsec\ in each coordinate.  The thin white contours underlying 
357: the MSX point source are the VLA 3.6~cm radio continuum contours from 
358: \citet{beu02b}, which have 0.7\arcsec\ resolution.  The coincidence 
359: of the continuum and \ntwo peaks in MMS2 suggests it is cold, 
360: and may be a pre-stellar massive star-forming core. 
361: The open ellipse in the lower left corner shows 
362: the size and orientation of the synthesized beam of the 3~mm continuum 
363: data.  
364: \label{fig:peaks}} 
365: \end{figure}
366: 
367: 
368: 
369: 	Based on C$^{18}$O observations, \citet{on96} derived a mean 
370: virial parameter of $\sim$0.6 for star-forming molecular cloud cores in 
371: Taurus.  Similarly, \citet{tach} derive a mean virial ratio of 0.9 for 
372: C$^{18}$O cores in $\rho$~Oph, with smaller ratios being associated with 
373: more actively star-forming cores.  The C$^{18}$O cores in Taurus and 
374: $\rho$~Oph have typical masses in the tens of solar masses and radii of 
375: $\sim$0.2~pc.  Despite the fact that these cores may have sufficient mass 
376: to form high-mass stars, the evidence from their environments is that they 
377: will instead form clusters of low-mass stars.  The principal 
378: distinguishing factor between the higher-mass cores in $\rho$~Oph and 
379: those in \iras\ seems to be that the \iras\ cores have much higher surface 
380: densities: $N_{\rm H_{2}} \sim 10^{21.5}$~\cmtwo\ in $\rho$~Oph 
381: \citep{tach} vs. $N_{\rm H_{2}} \sim 10^{24}$~\cmtwo\ in \iras.  The high 
382: surface densities of the \iras\ cores are typical of those found in the 
383: massive star-forming clumps surveyed in CS and C$^{34}$S by \citet{p97}.  
384: Thus, very dense gas seems to be key requirement for the formation of a 
385: massive star, as predicted by both the monolithic collapse and competitive 
386: accretion scenarios.
387: 
388: \subsection{Chemical Signatures of Evolutionary States}
389: \label{sec:chemistry}
390: 	
391: 	The rapid evolution of the physical conditions in massive 
392: star-forming objects drives substantial chemical evolution in these 
393: objects.  The time evolution of the abundances of particular 
394: molecules has not yet been well established for massive cores, but we can 
395: take some cues from the chemical studies of low-mass star-forming cores.  
396: In particular, we note that numerous studies have shown that, as a 
397: collapsing low-mass core heats up, CO desorption from grain mantles causes 
398: substantial destruction of \ntwo.  The destruction of \ntwo\ lowers the 
399: \ntwo\ abundance near the center of the core (e.g. 
400: \citealt{lee04,dif04,ber02}).  This drop in the \ntwo\ abundance has been 
401: observed in the centres of several low-mass cores, such as L483 
402: \citep{jor04} and Barnard 1c \citep{mat06}.  \citet{pir07} have shown 
403: that, in massive star-forming objects containing IRAS point sources, there 
404: is a measurable decrease in the \ntwo\ abundance toward the column density 
405: peaks, again likely indicative of \ntwo\ destruction in these evolved 
406: sources.
407: 
408: 	Although the sequence of chemical processes at work in high-mass 
409: star formation likely differs from that in low-mass star formation, the 
410: destruction of \ntwo\ by CO is likely to occur in both instances.  In 
411: \iras, the displacement between the dust and \ntwo\ emission peaks in 
412: MMS1, although small, suggests that the destruction of \ntwo\ near the 
413: massive protostar in MMS1 may already be underway.
414: Figure~\ref{fig:peaks} 
415: shows our 3~mm continuum image overlaid with the peak positions of five of 
416: the molecular species we observed, as well as the positions of the 
417: H$_{2}$O maser (taken from Table~2 of \citealt{beu02b}), the MSX point 
418: source, and contours of 3.6~cm VLA radio continuum 
419: emission\footnote{The position of the MSX point source was taken from the 
420: MSX point source catalog \citep{msxcat}.  The 3.6~cm radio continuum data 
421: were provided by H. Beuther.}.  The coincidence of the radio continuum and 
422: MSX emission with the dust peak in MMS1 suggests that a massive protostar 
423: is forming there.  The displacement between the dust and \ntwo\ peaks 
424: suggests that heating by the massive protostar may have caused the 
425: desorption of CO from dust grain mantles and the subsequent destruction of 
426: \ntwo\ near the MMS1 peak.  We emphasize, however, that the displacement 
427: between the \ntwo\ and dust peaks in MMS1 is less than a beam width, and 
428: may not be significant.  It is merely suggestive of chemical evolution in 
429: MMS1.
430: 
431: 	Similarly, we can interpret the close correspondence of the dust 
432: and \ntwo\ peaks in MMS2 as evidence that it is cold, chemically 
433: unevolved, and presumably therefore also very young.  Its mass and 
434: apparent lack of chemical evolution suggests that MMS2 is a good candidate 
435: pre-stellar massive star-forming object.
436: 
437: 	Our interpretation of MMS1 as chemically and dynamically more 
438: evolved than MMS2 is supported by the virial ratios of the two cores: 
439: lower in MMS1, indicating more gravitational binding, and higher in MMS2, 
440: indicating less binding and therefore perhaps an earlier evolutionary 
441: state (see \S~\ref{sec:morph}).
442: 
443: \section{Spectral Line Results: Kinematics}
444: \label{sec:spec}
445: 
446: \subsection{Outflows}
447: 
448: Several previous studies have demonstrated the presence of at least one 
449: outflow in \iras.  \citet{mz03} mapped the region in CO 3--2 using the 
450: 10~m Heinrich Hertz Telescope telescope and found overlapping red- and 
451: blue-shifted outflow lobes emanating from a point not precisely coincident 
452: with the IRAS point source.  Using BIMA array CO 2--1 observations, 
453: \citet{b04} estimated the mass entrained in this outflow to be 119~\msun.  
454: Among the lines we observed, three are known outflow tracers: \hcop\ 1--0, 
455: \meth\ 2--1, and SiO 2--1.  These three lines give strikingly different 
456: pictures of the outflow activity in the \iras\ region.
457: 
458: \subsubsection{\hcop and SiO}
459: \label{sec:hcopsio}
460: 
461: 	Figure~\ref{fig:hcop_outflow} shows contours of red- and 
462: blue-shifted \hcop\ from our BIMA array observations superimposed on the 
463: 3~mm continuum map.  This outflow clearly corresponds to the one seen by 
464: previous authors \citep{mz03,b04}.  The \hcop\ outflow is difficult to 
465: image because there is a significant background of \hcop\ associated with 
466: the underlying molecular cloud core, a source of confusion possibly 
467: amplified by the limited spatial resolution of our \hcop\ observations.  
468: In constructing Figure~\ref{fig:hcop_outflow}, we chose to display 
469: emission originating only in those velocity ranges which best highlight 
470: the outflow emission.  Curiously, the SiO 2--1 emission in the region does 
471: not appear to trace the same outflow as seen in \hcop, or at least not the 
472: same part of it.  Figure~\ref{fig:sio_outflow} shows the integrated 
473: intensity of SiO 2--1 over two velocity ranges: one centered at the 
474: systematic velocity and the other exactly matching that used to highlight 
475: the red lobe of the \hcop\ outflow in Figure~\ref{fig:hcop_outflow}.  We 
476: did not detect any SiO 2--1 emission at the 3\sig\ level over the range of 
477: velocities which define the blue lobe of the \hcop\ outflow.  The only 
478: suggestion of an association between the \hcop\ outflow and the SiO 
479: emission is that they are both extended and collinear.  However, the 
480: velocity structures of the emission in the two molecules preclude their 
481: interpretation as parts of the same outflow.  A more likely scenario is 
482: that the SiO peak at the systematic velocity is associated with hot and/or 
483: shocked gas in the immediate vicinity of the massive protostar.  
484: Figure~\ref{fig:sio_outflow} shows that the SiO emission peak at the 
485: systematic velocity is spatially coincident with the H$_{2}$O maser 
486: identified by \citet{beu02b}.  The two red-shifted SiO emission peaks to 
487: the northwest are harder to explain.  These peaks may represent the red 
488: lobe of a separate outflow from the one delineated by the \hcop\ emission, 
489: or they may be entirely unrelated to outflow activity.  If these SiO 
490: emission peaks are 
491: unrelated to outflow emission--if, say, they were associated with other 
492: protostars--it is difficult to explain why they do not appear in other 
493: tracers.  If the SiO peaks are associated with outflow activity, we expect 
494: that this outflow would be either weaker or chemically very different from 
495: the prominent outflow seen clearly in \hcop.  
496: 
497: \begin{figure}
498: \begin{center}
499: \includegraphics[width=0.8\columnwidth,angle=270]{f4.eps}
500: \caption{Integrated intensity of high-velocity \hcop\ 1--0 emission 
501: plotted over 
502: continuum emission (\emph{greyscale and thin solid contours}).  The blue 
503: outflow lobe (\emph{dotted contours}) is integrated over the velocity 
504: range $-61.2$ to $-63.3$ \kms\ and the red lobe (\emph{thick solid 
505: contours}) 
506: is integrated over the range $-44.8$ to $-46.9$ \kms.  Contours begin at 
507: 6~\sig\ for the red lobe and 4~\sig\ for the blue lobe, where \sig\ = 
508: 108~\mjbkms, and increase by steps of 1~\sig. The ellipse in the 
509: lower right shows the size and orientation of the \hcop\ beam.  The 
510: blue-lobe \hcop\ emission prominently overlaps MMS1, though their peaks do 
511: not coincide.  The overlap of the red and blue 
512: lobes is probably an indication that the outflow is strongly inclined to 
513: the plane 
514: of the sky.  \label{fig:hcop_outflow}}
515: \end{center} 
516: \end{figure}
517: 
518: \begin{figure}
519: \begin{center}
520: \includegraphics[width=0.8\columnwidth,angle=270]{f5.eps}
521: \caption{Integrated intensity of SiO~2--1 plotted over 3~mm continuum 
522: emission.  The thick grey contours show the intensity of SiO 2--1 
523: integrated over the same velocity range as the red lobe of the \hcop\ 1--0 
524: outflow, namely $-44.8$ to $-46.9$ \kms\ (see 
525: Fig.~\ref{fig:hcop_outflow}).  
526: The black contours show the intensity of SiO 2--1 around the 
527: systematic velocity, integrated from $-51$ to $-55$ \kms.  No SiO emission 
528: was detected over the velocity range corresponding to the blue lobe of 
529: the \hcop\ outflow.  The thin grey contours and grey scale represent the 
530: continuum as in Fig.~\ref{fig:hcop_outflow}.  All contours start at 3\sig\ and increase by 
531: steps of 1\sig, where \sig\ = 120~\mjbkms\ for the red lobe and \sig\ = 
532: 150~\mjbkms\ for the emission integrated around the systematic velocity.  
533: The 
534: white triangle indicates the 
535: position of the water 
536: maser, which is coincident with the local peak of the SiO emission.  The 
537: ellipse at the lower right 
538: shows the size and orientation of the SiO beam. \label{fig:sio_outflow}}
539: \end{center} 
540: \end{figure}
541: 
542: 
543: 	The identity of the driving source of the \hcop\ outflow is not 
544: clear.  There is no candidate protostellar source at the point where the 
545: two outflow lobes overlap.  The driving source is probably the MSX point 
546: source, which is spatially coincident with the peak of MMS1 and the 3.6~cm 
547: continuum emission.  This source is presumably the principal massive 
548: protostar in the region.  Less clear, however, is whether the H$_{2}$O 
549: maser and SiO emission are also associated with this massive protostar, or 
550: whether they trace another source entirely.  Just as the MSX source aligns 
551: neatly with the continuum peak of MMS1, the H$_{2}$O maser aligns neatly 
552: with the peak of the systematic-velocity SiO emission.  Neither pair of 
553: features coincides with the other.  Hence, we speculate that there may be 
554: two massive protostars within the MMS1 core, one situated at the 
555: millimeter continuum peak and one displaced a few arcseconds to the south.
556: 
557: 	There is no unambiguous evidence of an outflow emanating from 
558: MMS2.  The \hcop\ spectrum at the peak of MMS2 does show significant line 
559: wings (see Fig.~\ref{fig:infall_spectra}), but these could be caused by 
560: 'spill-over' from the \hcop\ outflow emanating from the vicinity of MMS1.  
561: As can be seen in Figure~\ref{fig:hcop_outflow}, some of the \hcop\ 
562: emission which appears to be associated with the MMS1 outflow also 
563: overlaps MMS2.  There is some evidence in the \hcop\ data 
564: cube of a second \hcop\ outflow in the region, but the spatial resolution 
565: of the data make it difficult to associate it conclusively with either 
566: MMS1 or MMS2.  To determine conclusively whether there is an outflow 
567: emanating from MMS2, we would require data with higher spatial resolution 
568: and preferably in a transition which traces outflows more selectively 
569: than does \hcop.
570: 
571: \subsubsection{\meth}
572: 
573: As in many other high-mass star-forming regions, the \meth\ 2--1 emission 
574: in \iras\ is very complex.  Figure~\ref{fig:meth_outflow} shows the total 
575: integrated intensity of \meth\ in \iras.  The \meth\ emission in \iras\ is 
576: distributed widely in velocity and, in general, does not correlate neatly 
577: in either velocity or position with other outflow features.  The exception 
578: is the elongated feature labeled `1' in Figure~\ref{fig:meth_outflow}, 
579: which is coincident both spatially and in velocity with the SiO peaks 
580: shown in Figure~\ref{fig:sio_outflow}.  A prominent bar of \meth\ emission 
581: overlaps most, but not all, of the continuum-emitting region.  The 
582: emission peak of the \meth\ bar does not coincide with either continuum 
583: peak.
584: 
585: 
586: 
587: \begin{figure} \begin{center} 
588: \includegraphics[width=0.8\columnwidth,angle=270]{f6.eps} 
589: \caption{Integrated intensity of \meth\ 2--1 from -40.4 to -67.7 \kms, 
590: plotted over 3~mm continuum emission.  The black contours show the 
591: intensity of \meth\ 2--1 integrated over the two brightest components of 
592: the line, starting from 3\sig\ and increasing by increments of 2\sig, 
593: where \sig\ = 355 \mjbkms.  
594: The thin grey contours and greyscale represent the continuum as in 
595: Fig.~\ref{fig:contmodel}.  The size and orientation of the \meth\ 
596: synthesized beam are indicated by the ellipse in the lower right.  The 
597: features labeled in the diagram are (1) a possible \meth\ counterpart to 
598: the SiO outflow, (2) a possible second outflow seen only in \meth, and 
599: (3) a conical structure which may represent the edges of a conical cavity 
600: carved out by a third outflow.  Both of the outflows suggested by 
601: features (2) and (3) appear to lie approximately in the plane of the 
602: sky.\label{fig:meth_outflow}} \end{center} \end{figure}
603: 
604: 
605: 
606: Although the \meth\ emission does not correlate well with the \hcop\ 
607: outflow, it can be interpreted as tracing several other possible outflows 
608: in the region.  For example, the dashed line labeled `2' in 
609: Figure~\ref{fig:meth_outflow} connects the \meth\ peak at the eastern edge 
610: of the image through a ridge of \meth\ emission to a second \meth\ peak 
611: near the western edge of the image.  In velocity, the \meth\ peaks at the 
612: two extremes of line `2' are separated by only a few \kms\, but their 
613: alignment suggests they may be part of an outflow lying nearly in the 
614: plane of the sky.  The lack of detected continuum emission at the position 
615: of either \meth\ peak supports this interpretation.  The driving source of 
616: this possible outflow is not evident.
617: 
618: 	The most prominent \meth\ emission feature is the large parabolic 
619: structure, labeled `3', which dominates the southern half of the image.  
620: This feature has no counterpart in any other tracer we observed.  Its 
621: velocity structure suggests it may be the conical cavity of an outflow 
622: lying approximately in the plane of the sky.  Again, the source of this 
623: possible outflow is not evident, although the alignment of the feature is 
624: consistent with a driving source within the MMS2 core.  If the driving 
625: source did lie within MMS2, the second outflow lobe could be confused with 
626: the elongated bar of emission at the systematic velocity traced by the 
627: continuum emission and in several of the lines.
628: 
629: \subsection{A ``Flattened Rotating Object''}
630: 
631: 	Four of the molecular transitions we observed have high critical 
632: densities and are known to trace dense star-forming gas.  These are the 
633: transitions for which integrated intensity maps are shown in 
634: Figure~\ref{fig:disklines}.  All four transitions trace similar structures 
635: in \iras: roughly elliptical distributions with long axes running from 
636: northwest to southeast, encompassing regions of high column density as 
637: traced by continuum emission (see Fig.~\ref{fig:contmodel}) The long axis 
638: of this elongated structure runs perpendicular to the axis of the 
639: prominent \hcop\ outflow.  A fit to the \htcop\ integrated intensity map 
640: shows that the position angle of the elongated feature is 
641: 35$^{\circ}\pm$5$^{\circ}$ degrees, while that of the \hcop\ outflow is 
642: approximately 130$^{\circ}$, meaning that the position angles of the two 
643: features differ by about 95$^{\circ}$.
644: 
645: 	As shown in Figure~\ref{fig:peaks}, the four dense gas tracers all 
646: peak at slightly different locations.  As we discussed in 
647: \S~\ref{sec:morph}, \ntwo\ and dust both trace the regions of highest 
648: column density.  Our \cs\ detection is too weak to determine the peak 
649: position accurately, except to say that it appears to correspond to the 
650: peak of the MMS1 core.  The peak of the \hcn\ emission matches that of the 
651: SiO emission and the position of the H$_{2}$O maser which, as we described 
652: in \S~\ref{sec:hcopsio}, is displaced by several arcseconds from the 
653: coincident dust continuum, radio continuum, and MSX peaks.  Notably, 
654: however, the maser and the emission peaks of \hcn\, SiO, and \cs\ all lie 
655: within one beam width of the MMS1 continuum peak, implying an association 
656: with MMS1.
657: 
658: 	The remaining dense gas tracer, \htcop, shows a distribution that 
659: requires some explanation.  \htcop\ peaks strongly at a point equidistant 
660: between the MMS1 and MMS2 continuum peaks, in a region of substantially 
661: lower column density as indicated by the dust emission.  The position of 
662: the \htcop\ peak does not match that of any other feature in the region 
663: (see Fig.~\ref{fig:peaks}).  It is unlikely that this peak is associated 
664: with any coherent physical structure; it is probably an artifact of 
665: depletion of \htcop\ onto dust grains at the column density peaks, 
666: leaving a remnant ``peak'' where depletion effects are least pronounced.  
667: Although \ntwo\ 1--0 and \htcop\ 1--0 peak at different locations, their 
668: kinematics clearly reveal that they trace the same contiguous structure in 
669: which the MMS1 and MMS2 peaks are embedded.  The axis ratio of the 
670: elongated structure seen in \htcop\ is 3.1.  Figure~\ref{fig:pvplot} shows 
671: position-velocity slices in \ntwoi\ and \htcop\ taken along the major axis 
672: of the elongated structure, at a position angle of 35$^{\circ}$.  As shown 
673: in the right panel of Figure~\ref{fig:pvplot}, both transitions exhibit 
674: the same velocity gradient, meaning that they trace the same large-scale 
675: structure.  Note that the signal-to-noise ratio of the \hcn\ and \cs\ data 
676: is too low to permit a meaningful analysis of their velocity structures.
677: 
678: \begin{figure*}
679: \begin{center}
680: \includegraphics[width=0.8\columnwidth,angle=270]{f7.eps}
681: \caption{Position-velocity slices taken along the long axis of the 
682: elongated structure seen in \ntwoi\ (\emph{left}) and \htcop\ 1--0 
683: (\emph{center}).  The position angle of the position-velocity cut is 
684: 35$^{\circ}$.  Contours begin at 3\sig\ and increase by intervals of 
685: 1\sig\ for \ntwoi\ and 2\sig\ for \htcop.  The white triangles in the 
686: left and center panels show the 
687: position-centroided velocities of the emission in each channel.  The right 
688: panel plots the two position-velocity curves on the same scale, 
689: illustrating the similarity between them.\label{fig:pvplot}}
690: \end{center} 
691: \end{figure*}
692: 
693: 
694: 	On large spatial scales, the elongated structure seen in 
695: Figure~\ref{fig:disklines} shares many characteristics in common with a 
696: stellar accretion disk.  For example, its long axis lies perpendicular to 
697: at least one outflow, there is a velocity gradient along its major axis, 
698: and it appears only in the continuum and emission from molecules which 
699: trace dense, non-outflowing gas.  However, the structure also has several 
700: characteristics which preclude its interpretation as an accretion disk, 
701: the first of which is its size.  The major axis of the \htcop\ emission 
702: over which there is a clear velocity gradient spans approximately 
703: 25\arcsec, or about 87,000~AU at 3.5~kpc.  By comparison, protostellar 
704: accretion disks around low-mass stars typically have radii in the range 
705: 50--500~AU \citep{dut96,kit02}.  Little is known about the plausible range 
706: of radii for accretion disks around massive protostars.  
707: Table~\ref{tab:bigdisks} gives an (incomplete) summary of the reported 
708: properties of candidate disks around massive protostars.  In most cases, 
709: the evidence that these objects are disks consists of three points: they 
710: are elongated in either the continuum or a dense gas tracer, the axis of 
711: elongation is approximately perpendicular to that of a detected outflow, 
712: and there is a velocity gradient along the long axis of the object 
713: observed in either molecular line or maser emission.  We note, however, 
714: that the range of reported physical properties of these disks is very 
715: large.  For example, their radii range over more than two orders of 
716: magnitude.  If we were to submit \iras\ as a candidate for this list based 
717: on its outflow properties and the shape and kinematics of its dense gas 
718: feature, it would only top out the list of an already very diverse sample.  
719: We note also that, while the molecular gas distribution in \htcop\ and 
720: possibly \ntwo\ could support the disk interpretation, the dust 
721: distribution does not.  The dust emission is better described as three 
722: strong peaks whereas an accretion disk should produce a centrally peaked 
723: distribution unless it had substantially fragmented.
724: 
725: \begin{deluxetable*}{lcccl}
726: \tablecaption{Candidate Massive Protostellar Disks\label{tab:bigdisks}}
727: \tablewidth{0pt}
728: \tablehead{\colhead{Object} & \colhead{Distance} & \colhead{Radius} & \colhead{Disk Mass} & \colhead{Ref.} \\ 
729: \colhead{} & \colhead{(kpc)} & \colhead{(AU)} & \colhead{(\msun)} & \colhead{} }
730: \startdata
731: G192.16-3.82 & 2 & 130 & 3 & \citet{shep01} \\
732: Cepheus A HW2 & 0.73 & 330 & 1--8 & \citet{patel05} \\
733: Becklin-Neugebauer Object & 0.46 & 800 & \nodata & \citet{jia05} \\
734: AFGL 5142 & 1.8 & 900 & \nodata & \citet{zha02} \\
735: IRAS 18089-1732 & 3.6 & $\sim$1000 & $\sim$16 & \citet{beu05} \\
736: NGC~7538 IRS 1N & 2.8 & 1000 & \nodata & \citet{pes04} \\
737: IRAS 20126+4104 & 1.7 & 1700 & 10 & \citet{ces97,ces99} \\
738: G24.78 A1 & 7.7 & 4000 & 23 & \citet{bel04} \\
739: G24.78 A2 & 7.7 & 4000 & 23 & \citet{bel04} \\
740: G24.78 C & 7.7 & 8000 & 23 & \citet{bel04} \\
741: G31.41 & 7.9 & 8000 & 23 & \citet{bel04} \\
742: NGC~7538~S & 2.8 & 15,000 & 400 & \citet{san03} \\
743: \enddata
744: \end{deluxetable*}
745: 
746: 	How, then, should we interpret the kinematics of \iras\ and, more 
747: generally, of very large structures with rotation signatures surrounding 
748: massive protostars?  We suggest that the observations to date, including 
749: those summarized in Table~\ref{tab:bigdisks}, are consistent with the 
750: hypothesis that massive protostars may be fed by Keplerian accretion disks 
751: (e.g. IRAS 20126+4104), but that these accretion disks are smoothly 
752: connected to much larger, flattened, rotating structures.  These 
753: ``flattened rotating objects'' are likely simply the remnants of the 
754: molecular cloud fragments from which the massive stars formed.  Existing 
755: observations, typically made with 
756: single interferometer configurations, only reveal rotational motions on 
757: the spatial scales visible with that interferometer configuration.  
758: Confirmation of the hypothesis that these rotating structures may extend 
759: from tens to thousands of AU would require multiple interferometric 
760: observations of each source with different interferometer configurations. 
761:  Such observations exist in only a few cases.  IRAS 20126+4104 is an 
762: instructive example: \citet{ces99} detect apparently Keplerian rotation on 
763: scales down to a few 100~AU using high-resolution CH$_{3}$CN observations 
764: while \citet{zha98} detect rotation out to radii of 5,000~AU in NH$_{3}$.  
765: We interpret our observations of \iras\ in a similar manner.  The 
766: optically thin \htcop\ line depicts a contiguous, flattened, 
767: slowly-rotating structure with a radius as large as 40,000~AU.  
768: \ntwo\ and dust continuum emission respectively reveal two and three 
769: condensations of higher column density within this apparently contiguous 
770: structure.  The cores detected in \ntwo\ and dust continuum emission may 
771: be fragmentation products of the original molecular cloud from which the 
772: \iras\ complex formed.
773: 
774: 
775: 
776: 
777: 	It is also possible that the velocity gradient seen in \htcop\ and 
778: \ntwo\ may simply indicate that the elongated structure is a sheared 
779: filament.  This hypothesis cannot be readily ruled out using the data in 
780: hand.
781: 
782: \subsection{Infall}
783: 
784: 	\citet{zhou92} showed that, in a dense molecular cloud 
785: core undergoing some form of inside-out collapse, optically thick lines 
786: will be skewed to the blue by an amount which depends on the infall 
787: velocity.  The combination of a double-peaked, asymmetrically blue \hcop\ 
788: line whose self-absorption peak is centered at the velocity of the 
789: optically thin \htcop\ line has been interpreted as an indicator of infall 
790: in both low-mass \citep{gre00} and high-mass \citep{ful05} star-forming 
791: cores.  Figure~\ref{fig:infall_spectra} shows spectra of \hcop\ and 
792: \htcop\ at the three continuum peaks seen in Figure~\ref{fig:contmodel}.  
793: At the continuum peaks of both MMS1 and MMS2, the \hcop\ line appears 
794: self-absorbed, indicating that it is optically thick.  At the position of 
795: MMS1, the line is slightly skewed to the red.  We note, however, that 
796: there is significant outflow contamination in the spectrum of the core at 
797: this position.  At the position of MMS2, the line profile is skewed to the 
798: blue, suggesting it may be collapsing.
799: 
800: \begin{figure}
801: \begin{center}
802: \includegraphics[width=1.3\columnwidth,angle=270]{f8.eps}
803: \caption{Spectra of \hcop\ (\emph{thick black lines}) and \htcop\ 
804: (\emph{thin black 
805: lines}) at the positions of MMS1 (\emph{top}), MMS2 peak (\emph{middle}), 
806: and MMS3
807: (\emph{bottom}).  The vertical line in each plot indicates the fitted peak 
808: of the \htcop\ spectrum, an indication of the systematic velocity at that 
809: position.
810: \label{fig:infall_spectra}}
811: \end{center} 
812: \end{figure}
813: 
814: 
815: 	Another possible interpretation of the double-peaked \hcop\ line 
816: profiles is that there are two sources along the line of sight.  This 
817: possibility would seem to be ruled out by the fact that the optically thin 
818: \htcop\ line is single-peaked throughout the region and it always peaks at 
819: the velocity of the \hcop\ self-absorption feature.  A third possible 
820: interpretation for the double-peaked and asymmetrically blue line profiles 
821: is that emission from the two cores is being blurred together in the 
822: comparatively large ($\sim10$\arcsec) \hcop\ beam.  To assess the 
823: likelihood of this possibility, we smoothed the \htcop\ data cube so that 
824: it had the same beam orientation and size as the \hcop\ beam.  Even in the 
825: smoothed data set, the \htcop\ line remains single-peaked at each 
826: continuum peak, thereby eliminating resolution effects as a likely 
827: explanation for the asymmetric line profiles.  Hence, we conclude that 
828: there is evidence for collapse in MMS2.
829: 
830: \section{Summary}
831: \label{sec:summary}
832: 
833: 	We used millimeter interferometry to study the structure and 
834: kinematics of the massive star-forming region IRAS 23033+5951.  We find 
835: that the region resolves into at least three distinct cores, at least two 
836: of which have sufficient mass to form one or more massive stars.  There is 
837: strong evidence from the spatial coincidence of strong radio continuum 
838: emission, an MSX point source, and a 3~mm continuum peak that MMS1 is 
839: forming a single massive star.  This core also powers at least one outflow 
840: seen in \hcop\ 1--0 emission and possibly a second outflow seen in SiO 
841: 2--1 and \meth\ 2--1.  The low virial ratio and moderately high density of 
842: MMS1 suggest it could be forming a massive star by collapse and disk 
843: accretion according to the ``monolithic collapse'' model.  The 
844: distribution of \ntwo\ in \iras\ strongly resembles that of the dust 
845: emission, suggesting that \ntwo\ is tracing the locations of highest 
846: column density.   Conversely, the fact that \htcop\ peaks between 
847: the column density peaks revealed by the dust and \ntwo\ emission probably 
848: reflects the depletion of \htcop\ onto grain mantles in the dense regions.  
849: 
850: 	MMS2 appears to be quiescent: it harbors no maser or MSX sources 
851: and has no unambiguous outflow signature.  The spatial coincidence of its 
852: \ntwo\ and dust continuum emission peaks suggest that it is a cold core in 
853: an early stage of evolution.  With an estimated total mass (dust and gas) 
854: of 205~\msun, a virial ratio of 0.9, and a possible spectral signature of 
855: collapse, MMS2 appears poised to form a massive star of its own.  It may 
856: be an example of a ``Class 0'' massive protostar.
857: 
858: 	Both of the cores described above are embedded in a smooth 
859: structure seen most clearly in the optically thin \htcop\ 1--0 line, but 
860: also seen in \hcn\ 1--0 and \hcop\ 1--0.  This structure has an axis ratio 
861: of 3.1, and its long axis lies almost exactly perpendicular to that of the 
862: \hcop\ outflow.  The \htcop\ emission peak of the structure is 
863: well-centered between the continuum peaks of the two bright cores.  There 
864: is a clear velocity gradient along the structure's long axis which is 
865: suggestive of rotation, although the size of the structure precludes its 
866: interpretation as a classical protostellar accretion disk.  We suggest it 
867: may either be the rotating, flattened remnant of the natal molecular cloud 
868: from which the two current cores formed, or simply a sheared filament.
869: 
870: 
871: \acknowledgements
872: 
873: 	M.~A. Reid is currently supported by a fellowship from the 
874: Harvard-Smithsonian Center for Astrophysics but, at the time the data for 
875: this paper were acquired, he was supported by a scholarship from the 
876: Natural Sciences and Engineering Research Council of Canada.  B.~C. 
877: Matthews is supported by the National Research Council of Canada.  The 
878: BIMA array was operated with support from the National Science Foundation 
879: under grants AST 02-28963 to UC Berkeley, AST 02-28953 to U. Illinois, and 
880: AST 02-28974 to U. Maryland. FCRAO is supported by NSF grant AST 02-28993. 
881:  The radio continuum data shown in Figure~\ref{fig:peaks} were generously 
882: provided by H. Beuther.
883: 
884: 
885: \begin{thebibliography}{}
886: \bibitem[Beltr\'{a}n et al.(2004)]{bel04}Beltr\'{a}n, M.~T., et al. 2004, \apj, 601, L187
887: \bibitem[Bergin et al.(2002)]{ber02}Bergin, E.~A., Alves, J., Huard, T., \& Lada, C.~J. 2002, \apj, 570, L101
888: \bibitem[Beuther et al.(2002a)]{beu02}Beuther, H., Schilke, P., Menten, K.~M., Motte, F., Sridharan, T.~K., \& Wyrowski, F., 2002a, \apj, 566, 945
889: \bibitem[Beuther et al.(2002b)]{beu02b}Beuther, H., Walsh, A., Schilke, P., Sridharan, T.~K., Menten, K.~M., \& Wyrowski, F. 2002b, \aap, 390, 289
890: \bibitem[Beuther et al.(2002c)]{beu02c}Beuther, H., Schilke, P., Sridharan, T.~K., Menten, K.~M., Walmsley, C.~M., \& Wyrowski, F.  2002c, \aap, 383, 892
891: \bibitem[Beuther, Schilke, \& Gueth(2004)]{b04}Beuther, H., Schilke, P., \& Gueth, F. 2004, \apj, 608, 330
892: \bibitem[Beuther et al.(2005)]{beu05}Beuther, H., Zhang, Q., Sridharan, T.~K., \& Chen, Y. 2005, \apj, 628, 800
893: \bibitem[Bonnell, Bate, \& Zinnecker(1998)]{bbz98}Bonnell, I.~A., Bate, M.~R., \& Zinnecker, H. 1998, \mnras, 298, 93
894: \bibitem[Bonnell et al.(2001a)]{b01a}Bonnell, I.~A., Bate, M.~R., Clarke, C.~J., \& Pringle, J.~E. 2001, \mnras, 323, 785
895: \bibitem[Bonnell et al.(2001b)]{b01b}Bonnell, I.~A., Clarke, C.~J., Bate, M.~R., \& Pringle, J.~E. 2001, \mnras, 324, 573
896: \bibitem[Bonnell, Vine, \&(Bate(2004)]{bvb04}Bonnell, I.~A., Vine, S.~G., \& Bate, M.~R. 2004, \mnras, 349, 735
897: \bibitem[Cesaroni et al.(1997)]{ces97}Cesaroni, R., Felli, M., Testi, L., Walmsley, C.~M., \& Olmi, L.  1997, \aap, 325, 725
898: \bibitem[Cesaroni et al.(1999)]{ces99}Cesaroni, R., Felli, M., Jenness, T., Neri, R., Olmi, L., Robberto, M., Testi, L., \& Walmsley, C. M. 1999, \aap, 345, 949
899: \bibitem[Di Francesco, Andr\'e, \& Myers(2004)]{dif04}Di Francesco, J., Andr\'e, P., \& Myers, P.~C. 2004, \apj, 617, 425
900: \bibitem[Dutrey et al.(1996)]{dut96}Dutrey, A. et al. 1996, \aap, 309, 493
901: \bibitem[Egan et al.(2003)]{msxcat}Egan, M.~P. et al., 2003, in Air Force Research Laboratory Technical Report AFRL-VS-TR-2003-1589
902: \bibitem[Williams, Fuller, \& Sridharan(2004)]{wil04}Williams, S.~J., Fuller, G.~A., \& Sridharan, T.~K. 2004, \aap, 417, 115
903: \bibitem[Fuller, Williams, \& Sridharan(2005)]{ful05}Fuller, G.~A., Williams, S.~J., \& Sridharan, T.~K. 2005, \aap, 442, 949
904: \bibitem[Gregersen \& Evans(2000)]{gre00}Gregersen, E.~M. \& Evans, N.~J. 2000, \apj, 538, 260
905: \bibitem[Harju, Walmsley, \& Wouterloot(1993)]{harju}Harju, J., Walmsley, C.~M., \& Wouterloot, J.~G.~A. 1993, A\&AS, 98, 51
906: \bibitem[Hildebrand(1983)]{h83}Hildebrand, R.~H. 1983, QJRAS, 24, 267
907: \bibitem[Jiang(2005)]{jia05}Jiang, Z., et al. 2005, \nat, 437, 112
908: \bibitem[J{\o}rgensen(2004)]{jor04}J{\o}rgensen, J.~K. 2004, \aap, 424, 589
909: \bibitem[Kitamura et al.(2002)]{kit02}Kitamura, Y. et al. 2002, \apj, 581, 357
910: \bibitem[Krumholz, McKee, \& Klein(2005)]{kmk05}Krumholz, M.~R., McKee, C.~F., \& Klein, R.~I. 2005, \nat, 438, 332
911: \bibitem[Krumholz(2006)]{k06}Krumholz, M.~R. 2006, \apj, 641, L45
912: \bibitem[Lee, Bergin \& Evans(2004)]{lee04}Lee, J.-E., Bergin, E.~A., \& Evans, N.~J. 2004, \apj, 617, 360
913: \bibitem[Mao \& Zeng(2003)]{mz03}Mao, R.~Q., \& Zeng, Q.  2003, AcASn, 44, 61
914: \bibitem[Matthews et al.(2006)]{mat06}Matthews, B.~C., Hogerheijde, M.~R., J{\o}rgensen, J.~K., \& Bergin, E.~A. 2006, \apj, 652, 1374
915: \bibitem[McKee \& Ostriker(2007)]{mo07}McKee, C.~F., \& Ostriker, E.~C.  
916: 2007, \araa, 45, 565
917: \bibitem[McKee \& Tan(2003)]{mt03}McKee, C.~F. \& Tan, J.~C. 2003, \apj, 585, 850
918: \bibitem[Onishi et al.(1996)]{on96}Onishi, T., Mizuno, A., Kawamura, A., Ogawa, H., \& Fukui, Y. 1996, \apj, 465, 815
919: \bibitem[Ossenkopf \& Henning(1994)]{oss94}Ossenkopf, V. \& Henning, Th. 1994, \aap, 291, 943
920: \bibitem[Patel et al.(2005)]{patel05}Patel, N., et al. 2005, \nat, 437, 109
921: \bibitem[Pestalozzi et al.(2004)]{pes04}Pestalozzi, M.~R., Elitzur, M., Conway, J.~E., \& Booth, R.~S. 2004, \apjl, 603, L113
922: \bibitem[Pirogov et al.(2007)]{pir07}Pirogov, L., Zinchenko, I., Caselli, P., \& Johansson, L.~E.~B.  2007, \aap, 461, 523
923: \bibitem[Plume et al.(1997)]{p97}Plume, R., Jaffe, D.~T., Evans, N.~J., Mart\'{i}n-Pintado, J., \& G\'{o}mez-Gonz\'{a}lez 1997, \apj, 476, 730
924: \bibitem[Sandell, Wright, \& Forster(2003)]{san03}Sandell, G., Wright, M., \& Forster, J.~R. 2003, \apj, 590, L45
925: \bibitem[Sault, Teuben, \& Wright(1995)]{miriad}Sault, R.~J., Teuben, P.~J., \& Wright, M.~C.~H.  1995, in ASP Conf. Ser. 77, Astronomical Data Analysis Software and Systems IV, ed. R.~A.~Shaw, H.~E.~Payne, \& J.~J.~E.~Hayes (San Francisco: ASP), 433
926: \bibitem[Shepherd, Claussen, \& Kurtz(2001)]{shep01}Shepherd, D.~S., Claussen, M.~J., \& Kurtz, S.~E. 2001, Science, 292, 1513
927: \bibitem[Sridharan et al.(2002)]{srid02}Sridharan, T.~K., Beuther, H., Schilke, P., Menten, K.~M., \& Wyrowski, F.  2002, \apj, 566, 931
928: \bibitem[Tachihara, Mizuno, \& Fukui(2000)]{tach}Tachihara, K., Mizuno, A., \& Fukui, Y. 2000, \apj, 528, 817
929: \bibitem[Zhang, Hunter, \& Sridharan(1998)]{zha98}Zhang, Q., Hunter, T.~R., \& Sridharan, T.~K. 1998, \apj, 505, L151
930: \bibitem[Zhang et al.(2002)]{zha02}Zhang, Q., Hunter, T.~R., Sridharan, T.~K., \& Ho, P.~T.~P. 2002, \apj, 566, 982
931: \bibitem[Zhou(1992)]{zhou92}Zhou, S. 1992, \apj, 394, 204
932: \end{thebibliography}
933: 
934: 
935: 
936: 
937: 
938: 
939: 
940: \end{document}
941: